Journhl of Medicinal Chemistry, 1975, Vol. 18, No. 4 399
CyclopentaVjisoquinoline Deriuatiues.2
E. Reich and I. H. Goldberg, Progr. Nucl. Acid Res. Mol. Biol., 3,183 (1964). (a) K. Alder, F. Pascher, and H. Vagt, Chem. Ber., 75, 1501 (1942):(b) J. A. Berson and G. B. Aspelin, Tetrahedron, 20, 2697 (i964). E. G. Wyss and 0. Schindler, Helu. Chim. Acta, 53, 1099 (1970). A. D. Eckard, A. Ledwith, and D. C. Sherrington, Polymer, 12,444 (1971). H. Rupe and F. Bernstein, Helu. Chim. Acta, 13,457 (1930). L. D. Taylor and R. B. Davis, J . Org. Chem., 28,1713(1963). A. R. Katritzky and J. M. Lagowski, Aduan. HeterocycL Chem., 1,311 (1963);2,l (1963). A. R. Katritzky and F. W. Maine, Tetrahedron, 20, 315 (1964). A. R.Katritzky, F. W. Maine, and S. Golding, Tetrahedron, 21,1693 (1965). D. A. Evans, G. F. Smith, and M. A. Wahid, J. Chem. Soc. B, 590 (1967). D. W. Jones, J. Chem. Soc. C , 1729 (1969). N. J. Mruk and H. Tieckelmann, Tetrahedron Lett., 14,1209 (1970). N. J. McCorkindale and A. W. McCulloch, Tetrahedron, 27, 4653 (1971). H. Plieninger, W. Muller, and K. Weinerth, Chem. Ber., 97, 667 (1964). W. A. Ayer, R. Hayatsu, P. de Mayo, S. T. Reid, and J. B. Stothers, Tetrahedron Lett., No. 18,648 (1961).
(30) L. A. Paquette and G. Slomp, J . Amer. Chem. Soc., 85,765 (1963). (31) E. C. Taylor and R. 0. Kan, J . Amer. Chem. Soc., 85, 776 (1963). (32) A. I. Meyers and P. Singh, J . Org. Chem., 35,3022 (1970). (33) J. N. Brown, R. L. R. Towns, and L. M. TreEonas, J . Amer. Chem. SOC.,93,7012 (1971). (34) W. E. Kreighbaum, W. F. Kavanaugh, W. T. Comer, and D. Deitchman, J. Med. Chem., 15,1131 (1972). (35) H. 0. House, “Modern Synthetic Reactions,” 2nd ed, W. A. Benjamin, Menlo Park, Calif., 1972,Chapter 9. (36) (a) J. Harley-Mason and T. J. Leeney, Proc. Chem. Soc., London, 368 (1964);(b) E. Bertele, H. Boos, J. D. Dunitz, F, Elsinger, A. Eschenmoster, I. Felner, H. P. Gribi, H. Gschwend, E. F. Meyer, M. Pesaro, and R. Scheffold, Angew. Chem., Int. Ed. Engl., 3, 490 (1964);(c) L.A. Paquette, J . Amer. Chem. SOC.,86,4096 (1964). (37) (a) E. A. Braude, A. G. Brook, and R. P. Linstead, J.Chem. SOC.,3569 (1954);(b) R. P. Linstead, E. A. Braude, L. M. Jackman, and A. N. Beames, Chem. Ind. (London), 1174 (1954);( c ) A. M. Creighton and L. M. Jackman, J. Chem. SOC.,3138 (1960). (38) D. E. Pearson, D. Cowan, and J. D. Beckler, J . Org. Chem., 24,504 (1959). (39) W. Baker, G.E. Coates, and F. Glockling, J . Chem. Soc., 1376 (1951). (40) M.Umeda and C. Heidelberger, Cancer Res., 28,2529 (1968). (41) F. Feist, Chem. Ber., 44,135 (1911).
Cyclopenta[flisoquinoline Derivatives Designed to Bind Specifically to Native Deoxyribonucleic Acid. 2. Synthesis of 6-Carbamylmethyl-8-methyl-7(5)H-cyclopenta[ flisoquinolin-3(2H)-onet and Its Interaction with Deoxyribonucleic Acids and Poly(deoxyribonuc1eotides)S Nitya G. Kundu, William Hallett, and Charles Heidelberger*pg McArdle Laboratory for Cancer Research, University of Wisconsin, Madison, Wisconsin 53706. Receiued July 19,1974
3-Ethoxy-8-methyl-5,6-dihydro-7H-cyclopenta~]isoquinolin-5-one (2)was converted to 6-carbethoxymethyl-3-ethoxy-8-methyl-5,6-dihydro-7H-cyclopenta~]isoquinolin-5-one (6)through an oxalyl derivative. Treatment of 6 with ammonia gave the corresponding amide 7 which on sodium borohydride reduction and subsequent dehydration yielded 6-carbamylmethyl-3-ethoxy-8-methyl-7(5)H-cyclopenta~]isoquinoline (9).The analogous ester 10 was similarly obtained from 6. Numerous attempts to dealkylate the 3-ethoxy group of 9 or 10 failed. However, 6 could easily be dealkylated on heating with 25% hydrochloric acid in a sealed tube. The ester, 6-carbethoxymethyl-8-methyl-5,6dihydro-7H-cyclopenta~]isoquinoline-3(2H),5-dione (ll),so obtained was converted to the corresponding amide 12 which on reduction with sodium borohydride and subsequent dehydration afforded the desired compound, 6-carbamylmethyl-8-methyl-7(5)H-cyclopenta~]isoquinolin-3-(2H)-one (1). 1 was found to be mildly cytotoxic against L5178Y mouse leukemia cells in culture. 1 was also found to bind to native calf thymus DNA. 1 inhibited RNA synthesis by a DNA-dependent RNA polymerase and a higher inhibition of RNA synthesis was observed when poly(dGdC) was used as a template than when poly(dA-dT) was used. A significant increase of thermal transition temperature of calf thymus DNA and poly(dG) poly(dC) was observed in the presence of 1. The accumulated evidence demonstrates that 1 interacts weakly with calf thymus DNA and interacts preferentially with poly(deoxyribonuc1eotides)containing GC pairs.
-
In the preceding paper,l we have described how the compound of structure 1 could bind to guanine-cytosine (GC) pairs of a DNA double helix by hydrogen bond formation. Starting from m- methyl-N-acetylbenzylaminewe have described a multistep synthesis of 3-ethoxy-8-methyl-5,6-dihydro-7H-cyclopenta~]isoquinolin-5-one (2) and the corresponding unsaturated compound 3 which constitute the basic skeleton of 1. In this paper we report the synthesis of compound 1 and its binding to DNA. In order to synthesize 1, it seemed preferable to put the side chain at the CS position on com+ The compound synthesized is a mixture of 7H and 5H isomers; only the structure corresponding to the 7H isomer is shown in the paper. t This work was supported in part by Grant CA-07175 from the National Cancer Institute, National Institutes of Health. 5 American Cancer Society Professor of Oncology.
CH2CONH2 1
2
3
pound 2, in the first place, followed by dealkylation. Attempts at the direct alkylation of 2 were n o t very success-
400 Journal ojMedicinal Chemistry, 1975, Vol. 18, No. 4
ful. Treatment of 2 with ethyl bromoacetate or iodoacetamide in the presence of potassium tert-butoxide2 in solvents such as DMF, THF, etc., yielded primarily the starting material, with some uncharacterizable water-soluble products. I t seems that under the conditions of direct alkylation, due to the low acidity of the methylene group adjacent to the carbonyl function, alkylation takes place predominantly on the nitrogen atom of the isoquinoline moiety, resulting in water-soluble quaternary salts. Similarly, alkylation could also be considered via the enamine3 of 2, but no enamine could be isolated from the reaction of 2 with pyrrolidine in the presence of p-toluenesulfonic acid in benzene. Whereas compound 2 could be easily formylated4 with ethyl formate in the presence of sodium ethoxide in benzene, the formyl derivative did not give any well-characterizable product from its reaction with iodoacetamide in the presence of sodium ethoxide. However, reaction of the formyl derivative of 2 with ethyl bromoacetate in the presence of potassium tert- butoxide in DMF led to compound 4 due to the 0-alkylation of the enolic form of the formyl derivative. The structure of 4 was established from its uv and nmr properties. In order to avoid 0-alkylation, 2 was converted to the corresponding oxalyl derivative 5 , in good yield, by treatment with diethyl oxalate in the presence of sodium ethoxide in benzene. The oxalyl derivative 5 reacted with ethyl bromoacetate in DMF in the presence of potassium tert- butoxide at 80’; alkylation and subsequent elimination of the oxalyl group took place. Removal of solvent and work-up yielded an oil which, after chromatography over Florisil [eluent benzene-methanol (10:1)], gave 6 as a crystalline solid. The keto ester 6 was converted to the corresponding amide 7 by heating in an autoclave with liquid ammonia. Alternatively, 6 could be converted by hydrolysis to the corresponding acid which was then converted to the acid chloride with thionyl chloride, and the acid chloride gave the amide by treatment with ammonia. Reduction of the keto amide 7 with sodium borohydride in ethanol furnished the hydroxyamide 8 which was dehydrated with p-toluenesulfonic acid in xylene to give the unsaturated amide 9. Similarly, reduction of 6 with sodium borohydride followed by treatment of the product with p -
Heidelber,qer, P t ai
toluenesulfonic acid in xylene afforded the unsaturated ester 10. Numerous attempts were made to dealkylate 6-carbamylmethyl-3-ethoxy-8-methyl-7(5)H-cyclopenta~]isoquinoline (9), the related ester 10, or the model compound 3. Use of vigorous reaction conditions, e.g., reflux with 48% hydrobromic a ~ i d ,usually ~ , ~ led to decomposition of the molecule or to polymeric products. Similarly, use of anhydrous aluminum c h l ~ r i d e , ~in, ~benzene, or boron tribromidegJO failed to give the desired dealkylated product. A much milder reagent, sodium iodide6 in acetic acid, is known to dealkylate alkoxypyrimidines but use of this reagent with 9 or 3 likewise failed to yield dealkylated products. An alternative approachl1,l2 was to heat the methiodides with lithium bromide in acetonitrile. However, when the methiodides of 3 or 9 were so treated, the corresponding N-methylisoquinolones could not be isolated. It is probable that the additional cyclopentane ring destabilizes the molecule to such an extent that the products break down under the reaction conditions. Finally, however, compound 6 was successfully dealkylated by heating with 25% hydrochloric acid in a sealed tube;I3J4 alternatively, 48% hydrobromic a ~ i dled ~ ? ~ to the same compound. Esterification of the dealkylated product afforded compound 11 which, on treatment with alcoholic ammonia, was converted to the corresponding amide 12.
11.
6
R
=
OC,H,
12, R = NH2
CH,CONH2 1
CH
2
CHOCH,C0,C2Hj 4
I
Reduction of 12 with sodium borohydride in ethanol and subsequent dehydration of the reduced product with hydrochloric acid in acetic acid yielded the desired final product, 1. Lactam-lactim tautomerism of the isoquinoline moiand its ety is known to occur in 3-hydro~yisoquinoline~~,~~ derivatives.l In the case of 3-hydroxycyclopenta[f]isoquinolines similar tautomerism is also observed. The lactam-lactim conversion was found to be solvent dependent. Uv spectral studies (Table I) indicate that a maximum at 357 nm can be ascribed to the lactim form, whereas the lactam form absorbs at ca. 420 nm. It is seen that 6-carbethoxymethyl-8-methyl-5,6-dihydro-7HcyclopentaV]isoquinoline-3(2H),5-dione (11) and the corresponding 6-carbamylmethyl derivative 12 exist exclusively in the lactam form in water, whereas in 95% ethanol, 11 and 12 are a mixture of the lactam and the lactim tautomers. 6-Carbamylmethyl-8-methyl-7(5)H-cyclopentav]isoquinolin-3(2 H )one (I), however, is predominantly in the lactam form both in water and 95% ethanol, as shown in Table I. The appearance of two aromatic methyl signals in the nmr shows that 1 is a mixture of 7H and 5H isomers. Cytotoxicity Studies. Preliminary studies with 1 a t M , according to the previously described assay procedure,” have shown it to be cytotoxic to L5178Y mouse leukemia cells in culture.
/pWIHj - pWjHa t
0 COCO,C,H, 5
CH2CONH,
8
CH,COR
0
6, R = OC,H, 7, R = NH,
CH,COR
9,R = NH, 10, R = (rC1H5
Journal of Medicinal Chemistry, 1975, Vol. 18, No. 4 401
CyclopentalflisoquinolineDerivatives. 2
Table I. Uv Spectra of Some Cyclopenta[f ]isoquinolines Compound
Solvent
6-Carbethoxymethyl-3-ethoxy-8- 95% ethanol methyl- 5,6-dihydro-7H-cyclopentalf]isoquinolin- 5-one (6) 6-Carbamylmethyl- 3- ethoxy- 89 5% ethanol methyl- 5,6-dihydro-7H-cyclopenta[f]isoquinolin- 5- one (7) 6-Carbethoxymethyl-3-ethoxy-8- 95% ethanol methyl- 7(5)H-cyclopenta[f]isoquinoline (IO) 6-Carbamylmethyl- 3-ethoxy- 895% ethanol methyl-7(5)H-cyclopenta[f]isoquinoline (9) 6-Carbethoxymethyl-8-methyl95% ethanol 5,6-dihydro-7H-cyclopentaWater [f]isoquinoline-3(2H)) ,5-dione (11) 6-Carbamylmethyl-%methyl95% ethanol 5,6-dihydro-7H-cyclopenta[f]- Water isoquinoline-3 (2H),5-dione (12) 6-Carbamylmethyl-8-methyl95% ethanol 7(5)~-cyclopenta[f]isoquinolin- Water 3(2H)-one (1)
X, nm (E)
357 (8360)
308 (5380)
296 (5340)
353 (7420)
307 (5780)
297 (5720)
357 (4580)
262 (1,150)
285 (1140)
233 (36,390)
357 (5450)
337 (2880)
295 (7060)
238 (4,460)
420 (3040) 415 (4770)
360 (2940)
325 (4000) 325 (5420)
312 (3830) 275 (4810)
239 (38,080) 238 (49,700)
420 (3900) 405 (5150)
360 (3660)
323 (4800) 317 (4870)
310 (4480) 275 (5170)
347 sh (1810) 348 sh (1600)
332 (2080) 333 (2000)
420 (1520) 418 (2500)
263 (15,300) 265 (20,000)
loor
I
a
f 80W 0
5I
60-
u
J
3
40-
2 i 0
20
I-
z W u 5
a
40
0 -10
40
30
50 60 70 TEMPERATURE ('C)
80
90
Figure 2. Thermal transition temperature of calf thymus DNA (0) and calf thymus DNA in the presence of 1 (A,0).The reaction conditions are described in the Experimental Section.
0 2
4
0
12
16
Conc. of GCB ( M )
20
x
IO4
Figure 1. Inhibition of DNA-dependent RNA polymerase reaction (E. coli) in the presence of 1. The templates used: calf thymus
-
DNA 0 , poly(dA-dT) .poly(dA-dT) 0,poly(dA) poly(dT) A, poly(dG) . poly(dC) 0,poly(dG-dC) poly(dG-dC) X. The reaction conditions are described in the Experimental Section. GCB (abbreviation for GC binder) refers to compound 1.
.
Table 11. Binding Studies with Calf Thymus DNA as Determined by Difference Spectroscopy ~~
Compound
*OD400 -0.031
Interaction with DNA's and Poly(deoxyribonuc1eotides). The cyclopenta[f]isoquinoline derivative 1 interacts with DNA's and poly(deoxyribonuc1eotides). This is evident from the effect of l on the DNA-dependent RNA polymerase (Escherichia coli) reaction as is shown in Figure 1. A higher inhibition of RNA synthesis was observed when poly(deoxyguanine-deoxycytosine) [poly(dG-dC)] was used as a template than when poly(deoxyadeninedeoxythymine) [poly(dA-dT)] was used. The data from the inhibition of the RNA-polymerase reaction are in good agreement with those we obtained from difference spectra and equilibrium dialysis experiments.ls Additional difference spectra studies have revealed the importance of the tricyclic moiety and the carbamylmethyl side chain for the binding of 1 with calf thymus DNA. As is seen in Table 11,
CH,CONH,
0.00
'
"
P
o
d0,Et
no interaction between 6-carbethoxymethyl-7-methylisoquinolin-3(2H)-one and calf thymus DNA could be observed. The interaction between 1 and calf thymus DNA is also demonstrated by the effect of 1 on the thermal transition temperature (T,) of native calf thymus DNA (Figure 2). A small but significant increase in T , of 3 . 5 O was observed.
402 Journal of Medicinal ChemcstQi, 1975, Vol 18, No 1
More significantly, in the presence of 1 although a small increase in T , of 3' was observed for poly(dG) poly(dC), no such change of T , was seen in the case of poly(dG-dC), poly(dA-dT), and poly(dA) poly(dT). From the above studies, it can be concluded that our predictions based on molecular models are fulfilled to a considerable extent. I t is evident that 1 definitely interacts with DNA's and polydeoxyribonucleotides although the magnitude of the interaction is small. This agrees well with our hydrogen bonded model since only two hydrogen bonds are formed per molecule of 1 in its interaction with DNA's and poly(deoxyribonuc1eotides). Also, in certain cases, 1 binds preferentially to poly(deoxyribonuc1eotides) containing GC pairs. In conclusion, it can be said that although alternative mechanisms of interaction are possible for the interaction of 1 with DNA's and poly(deoxyribonucleotides), the experiments we have reported so far favor our hydrogen bonded model. The effect of salt concentration on the stability of the complex between 1 and calf thymus DNA as evidenced by difference spectra results,ls the failure to isolate a complex by gel filtration, and the absence of any effect on the circular dichroism spectra of calf thymus DNA in the presence of 1 also support the hydrogen bonded model and disfavor any intercalation model for the binding of 1 with DNA's and poly(deoxyribonuc1eotides).
cyclopenta[f]isoquinolin-5-one (6). Sodium ethoxide (made from 230 mg, 10 mmo! of sodium) was suspended in 100 ml of benzene, the oxalyl derivative 5 (2.3 g, 6.74 mmol) was added, and the mixture was stirred under Nz for 2 hr. Benzene was removed in uacuo. The sodium salt was suspended in 50 ml of DMF, and ethyl bromoacetate (1.4 g, 8.38 mmol) in 20 ml of DMF was added dropwise with stirring at room temperature. The mixture was then heated at 80' with stirring for 1 2 hr, DMF was removed in L'QCUO, and the residue was treated with HzO and extracted with CHC1:3. The CHC13 extract was dried and evaporated to yield an oil, which was chromatographed over Florisil (60-100 mesh) using benzene and benzene-MeOH (10:l i as eluent. The benzene-MeOH fractions were combined and evaporated to yield an oil (2 g, 6.11 mmol. 90%, which was crystallized from ethanol as a light yellow 357 (8360), solid: mp 102-103'; P max 1735, 1720, 1690 cm-I; A,, 308 (53801, 296 (5340), 262 (1150); nmr (CDC13) 1.24 (t, 3 H , .I = ': Hz), 1.46 it, 3 H, J = 7 Hz). 2.42 (s, 2 H). 4.15 and 4.45 (two over lapping quartets, 4 H), 7.72 (s, 1 HI. 8.1 (s. 1 H), 8.81 (9, 1 H) Anal. (C1gHz1N04) C, H, N. 6-Carbamylmethyl-3-ethoxy-8-methyl-5,6-tihydro-7H-cyclopenta[f]isoquinolin-5-one (7). Method A. The keto ester 6 (500 mg. 1.53 mmol) was treated with liquid 3" at 132' (pressure 1500 lb) for 24 hr: NH:, was then evaporated and the residue was chromatographed on Florisil. The first eluents (acetone-cyclohexane: 1:l) yielded the unreacted ester. Further elution with benzene-MeOH (3:l) afforded the amide 7 (100 mg, 0.34 mmol, 22%), which crystallized from EtOH as small white needles: mp 239353 (7420). 307 (57801, 297 240'; u max 1685, 1630 cm-I; ,,,A C, H, N. 15720). Anal. (CI:H~*N;:O~) Method R. The keto ester 6 1100 mg, 0.3 mmoli was refluxed Experimental Section with 2 ml of concentrated HCI for 7 hr. After cooling overnight in Melting points were determined on a Thomas-Hoover melting the refrigerator, 80 mg (0.24 mmol) of yellow crystals of the keto point apparatus and are uncorrected. The uv spectra were recordacid was obtained: mp 205-206O. The keto acid (40 mg, 0.13 mmol) was allowed to stand at room temperature with SOClz ( 2 ml) for 10 ed on a Beckman DB-Gand quantitative measurements done on a hr. and SOCl;: was then removed in Lwcuo The residue was added Gilford 2400-S. Spectra were usually taken in 95% ethanol unless in benzene suspension to ether-benzene \1:1) saturated with "3. otherwise mentioned. The ir spectra were taken on a Beckman IRThe reaction was stirred at room temperature for 16 hr and then 10 as KBr plates. Nmr spectra (reported in 6) were recorded on a Solvent was refluxed for 2 hr with occasional saturation with "3. Perkin-Elmer R-12 in deuteriochloroform, using tetramethylsilane removed to yield a solid (14 mg, 0.05 mmol, 38%), which crystalas internal reference. Elemental analyses were performed by Gallized frc;m EtOH as white needles: mp and mmp (with a sample braith Laboratories, Inc., Knoxville, Tenn., or Spang Microanalytprepared by method Ai 239-240'. cal Laboratory, Ann Arbor, Mich. All analytical results were within 6-Carbamylmethyl-3-ethoxy-~-h~droxy-8-methyl-~,6-dihy0.4% of the theoretical values. dro-7H-cyclopenta[f]isoquinoline(8). A suspension of 7 (75 mg. 6-(Carbethoxymethylyloxy)methylene-3-ethoxy-8-methyl0.25 mmol) and NaB& (100 mg, 2.64 mmol) in EtOH (25 ml) was 5,6-dihydro-7H-cyclopenta[f]isoquinolin-5-one(4). To a wellstirred at room temperature for 4 hr. EtOH was removed in i;acuo stirred suspension of NaOMe (prepared from Na, 120 mg, 5.22 and the residue treated with H20 and filtered to yield a white solid mmol) in dry benzene (2 ml), ethyl formate (400 mg, 5.40 mmol) in (70 mg, 0.25 mmol. 92%) which was crystallized from EtOH: mp 2 ml of benzene was added, and the mixture was stirred for 0.5 hr. 340 (3810). 282 (4110), 272 188-189'; max 3360, 1670 cm-':, , ,A The reaction was cooled in an ice bath and compound 2 (105 mg, (46001. ilnal. (CI;H2oN&(, C. H , N. 0.44 mmol) was added in 15 ml of benzene. The whole mixture was 6-Carbamylmethyl-3-ethoxy-8-methyl-7(~)H-cyclopentastirred in an ice bath for 15 min and at room temperature for 45 [flisoquinoline (9). A mixture of 8 (25 mg, 0.08 mmol) and p-tolumin and was then decomposed with a phosphate buffer (0.05 M , enesulfonic acid (5 mg. 29 umol) was refluxed in 1 5 ml of xylene pH 8 ) and extracted wiih CHCb. After removal of solvent, 90 mg under N 2 . Xylene was removed in cacuo and the residue was treat(0.25 mmol, 57%) of the formyl derivative, mp 150-152', was obed with 5 ml of 2 i\' NaOH solution and extracted with CHCl:
